Electronic Journal of Differential Equations, Vol. 2006(2006), No. 117, pp. 1–11. ISSN: 1072-6691. URL: http://ejde.math.txstate.edu or http://ejde.math.unt.edu ftp ejde.math.txstate.edu (login: ftp) EXISTENCE AND UNIQUENESS OF PERIODIC SOLUTIONS FOR FIRST-ORDER NEUTRAL FUNCTIONAL DIFFERENTIAL EQUATIONS WITH TWO DEVIATING ARGUMENTS JINSONG XIAO, BINGWEN LIU Abstract. In this paper, we use the coincidence degree theory to establish the existence and uniqueness of T -periodic solutions for the first-order neutral functional differential equation, with two deviating arguments, (x(t) + Bx(t − δ))0 = g1 (t, x(t − τ1 (t))) + g2 (t, x(t − τ2 (t))) + p(t). 1. Introduction Consider the first-order neutral functional differential equation (NFDE), with two deviating arguments, (x(t) + Bx(t − δ))0 = g1 (t, x(t − τ1 (t))) + g2 (t, x(t − τ2 (t))) + p(t), (1.1) where τ1 , τ2 , p : R → R and g1 , g2 : R × R → R are continuous functions, B and δ are constants, τ1 , τ2 and p are T -periodic, g1 and g2 are T -periodic in the first argument, |B| = 6 1 and T > 0. The above equation has been used for the study of distributed networks containing lossless transmission lines [6, 7]. Hence, in recent years, the problem of the existence of periodic solutions for (1.1) has been extensively studied. For more details, we refer the reader to [1, 2, 4, 5, 6, 7, 9, 12] and the references cited therein. However, to the best of our knowledge, there exist no results for the existence and uniqueness of periodic solutions of (1.1). The main purpose of this paper is to establish sufficient conditions for the existence and uniqueness of T -periodic solutions of (1.1). The results of this paper are new and they complement previously known results. An illustrative example is given in Section 4. For ease of exposition, throughout this paper we will adopt the following notation: Z T 1/k |x|k = |x(t)|k dt , |x|∞ = max |x(t)|. t∈[0,T ] 0 2000 Mathematics Subject Classification. 34C25, 34D40. Key words and phrases. First order; neutral; functional differential equations; deviating argument; periodic solutions; coincidence degree. c 2006 Texas State University - San Marcos. Submitted March 16, 2006. Published September 26, 2006. Supported by grants 10371034 from the NNSF of China and 05JJ40009 from the Hunan Provincial Natural Science Foundation of China. 1 2 J. XIAO,B. LIU EJDE-2006/117 Let X = {x|x ∈ C(R, R), x(t + T ) = x(t), for all t ∈ R} be a Banach space with the norm kxkX = |x|∞ . Define the two linear operators A : X → X, (Ax)(t) = x(t) + Bx(t − δ); L : D(L) ⊂ X → X, Lx = (Ax)0 , (1.2) where D(L) = {x|x ∈ X, x0 ∈ C(R, R)}. We also define the nonlinear operator N : X → X by N x = g1 (t, x(t − τ1 (t))) + g2 (t, x(t − τ2 (t))) + p(t). By Hale’s terminology [4], a solution u(t) of (1.1) is that u ∈ C(R, R) such that Au ∈ C 1 (R, R) and (1.1) is satisfied on R. In general, u 6∈ C 1 (R, R). But from [9, Lemma 1], in view of |B| 6= 1, it is easy to see that (Ax)0 = Ax0 . So a T -periodic solution u(t) of (1.1) must be such that u ∈ C 1 (R, R). Meanwhile, according to [9, RT Lemma 1], we can easily get that ker L = R, and Im L = {x ∈ X : 0 x(s)ds = 0}. Therefore, the operator L is a Fredholm operator with index zero. Define the continuous projectors P : X → ker L and Q : X → X/ImL by setting Z 1 T x(s)ds, P x(t) = T 0 Z 1 T Qx(t) = x(s)ds. T 0 Hence, Im P = ker L and ker Q = Im L. Set LP = L|D(L)∩KerP , then LP has continuous inverse L−1 P defined by Z t Z T −1 1 L−1 y(t) = A sy(s)ds + y(s)ds . (1.3) P T 0 0 Therefore, it is easy to see from (1) and (1.3) that N is L-compact on Ω, where Ω is an open bounded set in X. 2. Preliminary Results In view of (1.2) and (1), the operator equation Lx = λN x is equivalent to the equation x0 (t) + Bx0 (t − δ) = λ[g1 (t, x(t − τ1 (t))) + g2 (t, x(t − τ2 (t))) + p(t)], where λ ∈ (0, 1). For convenience of use, we introduce the Continuation Theorem [2] as follows. Lemma 2.1. Let X be a Banach space. Suppose that L : D(L) ⊂ X → X is a Fredholm operator with index zero and N : Ω → X is L-compact on Ω, where Ω is an open bounded subset of X. Moreover, assume that all the following conditions are satisfied: (1) Lx 6= λN x, for all x ∈ ∂Ω ∩ D(L), λ ∈ (0, 1); (2) N x 6∈ ImL, for all x ∈ ∂Ω ∩ ker L; (3) For the Brower degree, deg{QN, Ω ∩ ker L, 0} = 6 0. Then the equation Lx = N x has at least one solution on Ω ∩ D(L). EJDE-2006/117 EXISTENCE AND UNIQUENESS OF PERIODIC SOLUTIONS 3 By using a similar argument of the proof of [8, Lemma 2.5], from [3, Theorem 225], we can obtain the following Lemma. Lemma 2.2. Let x(t) ∈ X ∩ C 1 (R, R) . Suppose that there exists a constant D ≥ 0 such that |x(τ0 )| ≤ D, τ0 ∈ [0, T ]. Then √ T |x|2 ≤ |x0 |2 + T D. (2.1) π Lemma 2.3 ([10]). Let µ ∈ [0, T ] be a constant, δ ∈ C(R, R) be periodic with period T , and supt∈[0,T ] |δ(t)| ≤ µ. Then for any h ∈ C 1 (R, R) which is periodic with period T , we have Z T Z T 2 2 |h(s) − h(s − δ(s))| ds ≤ 2µ |h0 (s)|2 ds. (2.2) 0 0 For the next lemma we need the following conditions (H) For i = 1, 2, there exist a constants µi and an integers Ki such that µi = sup |τi (t) − Ki T | ≤ T. t∈[0,T ] (A0) One of the following conditions holds: (1) (gi (t, u1 ) − gi (t, u2 ))(u1 − u2 ) > 0, for i = 1, 2, ui ∈ R, for all t ∈ R and u1 6= u2 ; (2) (gi (t, u1 ) − gi (t, u2 ))(u1 − u2 ) < 0, for i = 1, 2, ui ∈ R, for all t ∈ R and u1 6= u2 ; (A0’) One of the following conditions holds: √ √ (1) there exists constants b1 and b2 such that b1 ( 2µ1 + Tπ )+b2 ( 2µ2 + Tπ ) < 1 − |B|, and |gi (t, u1 ) − gi (t, u2 )| ≤ bi |u1 − u2 |, for i = 1, 2, ui ∈ R, ∀t ∈ R, √ √ (2) There exists constants b1 and b2 such that b1 ( 2µ1 + Tπ ) + b2 ( 2µ2 + T π ) < |B| − 1, and |gi (t, u1 ) − gi (t, u2 )| ≤ bi |u1 − u2 |, for i = 1, 2, ui ∈ R, ∀t ∈ R. Lemma 2.4. Under assumptions (A0) and (A0’), Equation (1.1) has at most one T -periodic solution. Proof. Suppose that x1 (t) and x2 (t) are two T -periodic solutions of (1.1). Then (x1 (t) + Bx1 (t − δ))0 − g1 (t, x1 (t − τ1 (t))) − g2 (t, x1 (t − τ2 (t))) = p(t) and (x2 (t) + Bx2 (t − δ))0 − g1 (t, x2 (t − τ1 (t))) − g2 (t, x2 (t − τ2 (t))) = p(t). This implies [(x1 (t) − x2 (t)) + B(x1 (t − δ) − x2 (t − δ))]0 − (g1 (t, x1 (t − τ1 (t))) − g1 (t, x2 (t − τ1 (t)))) − (g2 (t, x1 (t − τ2 (t))) − g2 (t, x2 (t − τ2 (t)))) = 0. (2.3) Set Z(t) = x1 (t) − x2 (t). Then, from (2.3), we obtain Z 0 (t) + BZ 0 (t − δ) − (g1 (t, x1 (t − τ1 (t))) − g1 (t, x2 (t − τ1 (t)))) − (g2 (t, x1 (t − τ2 (t))) − g2 (t, x2 (t − τ2 (t)))) = 0. (2.4) 4 J. XIAO,B. LIU EJDE-2006/117 Thus, integrating (2.4) from 0 to T , we have Z T [(g1 (t, x1 (t − τ1 (t))) − g1 (t, x2 (t − τ1 (t)))) 0 + (g2 (t, x1 (t − τ2 (t))) − g2 (t, x2 (t − τ2 (t))))]dt = 0. Therefore, in view of integral mean value theorem, it follows that there exists a constant γ ∈ [0, T ] such that (g1 (γ, x1 (γ − τ1 (γ))) − g1 (γ, x2 (γ − τ1 (γ)))) + (g2 (γ, x1 (γ − τ2 (γ))) − g2 (γ, x2 (γ − τ2 (γ)))) = 0. (2.5) From (A0), (2.5) implies (x1 (γ − τ1 (γ)) − x2 (γ − τ1 (γ)))(x1 (γ − τ2 (γ)) − x2 (γ − τ2 (γ))) ≤ 0. Since Z(t) = x1 (t) − x2 (t) is a continuous function on R, it follows that there exists a constant ξ ∈ R such that Z(ξ) = 0. (2.6) Let ξ = nT + γ e, where γ e ∈ [0, T ] and n is an integer. Then, (2.6) implies that there exists a constant γ e ∈ [0, T ] such that Z(e γ ) = Z(ξ) = 0. (2.7) Then, from Lemma 2.2, using Schwarz inequality and the inequality Z t Z T |Z(t)| = |Z(e γ) + Z 0 (s)ds| ≤ |Z 0 (s)|ds, for all t ∈ [0, T ], γ e we obtain |Z|∞ ≤ 0 √ T |Z 0 |2 , and |Z|2 ≤ T 0 |Z |2 . π (2.8) Now, we consider two cases. Case (i). If (A0’)(1) holds, multiplying both sides of (2.4) by Z 0 (t) and then integrating them from 0 to T , using (H), (2.2), (2.8) and Schwarz inequality, we have |Z 0 |22 Z = T |Z 0 (t)|2 dt 0 T Z 0 = −B Z 0 Z Z (t)Z (t − δ)dt + 0 T T (g1 (t, x1 (t − τ1 (t))) − g1 (t, x2 (t − τ1 (t))))Z 0 (t)dt 0 (g2 (t, x1 (t − τ2 (t))) − g2 (t, x2 (t − τ2 (t))))Z 0 (t)dt + 0 ≤ |B||Z 0 |22 + b1 Z T |x1 (t − τ1 (t)) − x2 (t − τ1 (t))||Z 0 (t)|dt 0 Z T |x1 (t − τ2 (t)) − x2 (t − τ2 (t))||Z 0 (t)|dt + b2 0 ≤ |B||Z 0 |22 Z + b1 T |Z(t − τ1 (t)) − Z(t)||Z 0 (t)|dt + b1 0 Z + b2 0 T |Z(t − τ2 (t)) − Z(t)||Z 0 (t)|dt + b2 Z T |Z(t)||Z 0 (t)|dt 0 Z 0 T |Z(t)||Z 0 (t)|dt EJDE-2006/117 EXISTENCE AND UNIQUENESS OF PERIODIC SOLUTIONS T Z ≤ |B||Z 0 |22 + b1 ( 5 1 |Z(t − τ1 (t)) − Z(t)|2 dt) 2 |Z 0 |2 + b1 |Z|2 |Z 0 |2 0 T Z + b2 ( 1 |Z(t − τ2 (t)) − Z(t)|2 dt) 2 |Z 0 |2 + b2 |Z|2 |Z 0 |2 0 T Z = |B||Z 0 |22 + b1 ( 1 |Z(t − (τ1 (t) − K1 T )) − Z(t)|2 dt) 2 |Z 0 |2 + b1 |Z|2 |Z 0 |2 0 T Z + b2 ( 1 |Z(t − (τ2 (t) − K2 T )) − Z(t)|2 dt) 2 |Z 0 |2 + b2 |Z|2 |Z 0 |2 0 √ √ T T ≤ [|B| + b1 ( 2µ1 + ) + b2 ( 2µ2 + )]|Z 0 |22 . π π From (2.8) and (A0’)(1), the above inequalit implies Z(t) ≡ Z 0 (t) ≡ 0, for all t ∈ R. Hence, x1 (t) ≡ x2 (t), for all t ∈ R. Therefore, (1.1) has at most one T -periodic solution. Case (ii). If (A0’)(2) holds, multiplying both sides of (2.4) by Z 0 (t − δ) and then integrating them from 0 to T , using (H), (2.2), (2.8) and Schwarz inequality, we have |B||Z 0 |22 Z T =| B|Z 0 (t − δ)|2 dt| 0 T Z Z 0 (t)Z 0 (t − δ)dt =|− Z 0 T (g1 (t, x1 (t − τ1 (t))) − g2 (t, x2 (t − τ1 (t))))Z 0 (t − δ)dt + 0 Z T (g2 (t, x1 (t − τ2 (t))) − g2 (t, x2 (t − τ2 (t))))Z 0 (t − δ)dt| + 0 ≤ |Z 0 |22 + b1 T Z |x1 (t − τ1 (t)) − x2 (t − τ1 (t))||Z 0 (t − δ)|dt 0 T Z |x1 (t − τ2 (t)) − x2 (t − τ2 (t))||Z 0 (t − δ)|dt + b2 0 ≤ |Z 0 |22 T Z Z 0 |Z(t − τ1 (t)) − Z(t)||Z (t − δ)|dt + b1 + b1 0 |Z(t)||Z 0 (t − δ)|dt 0 T Z |Z(t − τ2 (t)) − Z(t)||Z 0 (t − δ)|dt + b2 + b2 T 0 T Z |Z(t)||Z 0 (t − δ)|dt 0 Z ≤ |Z 0 |22 + b1 ( T 1 |Z(t − τ1 (t)) − Z(t)|2 dt) 2 |Z 0 |2 + b1 |Z|2 |Z 0 |2 0 Z + b2 ( T 1 |Z(t − τ2 (t)) − Z(t)|2 dt) 2 |Z 0 |2 + b2 |Z|2 |Z 0 |2 0 = |Z 0 |22 Z + b1 ( 0 T 1 |Z(t − (τ1 (t) − K1 T )) − Z(t)|2 dt) 2 |Z 0 |2 + b1 |Z|2 |Z 0 |2 6 J. XIAO,B. LIU Z + b2 ( EJDE-2006/117 T 1 |Z(t − (τ2 (t) − K2 T )) − Z(t)|2 dt) 2 |Z 0 |2 + b2 |Z|2 |Z 0 |2 0 √ √ T T ≤ [1 + b1 ( 2µ1 + ) + b2 ( 2µ2 + )]|Z 0 |22 π π Then using the methods similar to those used in Case (i), from the above inequality, (2.8), and (A0’)(2), we can conclude that (1.1) has at most one T -periodic solution. The proof of Lemma 2.4 is now complete. For the next lemma we use the following assumptions: (A1) x(g1 (t, x) + g2 (t, x) + p(t)) > 0, for all t ∈ R, |x| ≥ d; (A2) x(g1 (t, x) + g2 (t, x) + p(t)) < 0, for all t ∈ R, |x| ≥ d. Lemma 2.5. Assume (A0) and that there exists a positive constant d such that one of the two conditions (A1) or (A2) holds. If x(t) is a T -periodic solution of (2), then √ (2.9) |x|∞ ≤ d + T |x0 |2 . Proof. Let x(t) be a T -periodic solution of (2). Then, integrating (2) from 0 to T , we have Z T [g1 (t, x(t − τ1 (t))) + g2 (t, x(t − τ2 (t))) + p(t)]dt = 0. 0 This implies that there exists a constant t1 ∈ R such that g1 (t1 , x(t1 − τ1 (t1 ))) + g2 (t1 , x(t1 − τ2 (t1 ))) + p(t1 ) = 0. (2.10) We show next the following Claim: If x(t) is a T -periodic solution of (2), then there exists a constant t2 ∈ R such that |x(t2 )| ≤ d. (2.11) Assume, by way of contradiction, that (2.11) does not hold. Then |x(t)| > d, for all t ∈ R, which, together with (A1), (A2) and (2.10), implies that one of the following relations holds: x(t1 − τ1 (t1 )) > x(t1 − τ2 (t1 )) > d; (2.12) x(t1 − τ2 (t1 )) > x(t1 − τ1 (t1 )) > d; (2.13) x(t1 − τ1 (t1 )) < x(t1 − τ2 (t1 )) < −d; (2.14) x(t1 − τ2 (t1 )) < x(t1 − τ1 (t1 )) < −d. (2.15) If (2.12) holds, in view of (A0)(1), (A0)(2), (A1) and (A2), we shall consider four cases as follows. Case (i). If (A1) and (A0)(1) hold, according to (2.12), we obtain 0 < g1 (t1 , x(t1 − τ2 (t1 ))) + g2 (t1 , x(t1 − τ2 (t1 ))) + p(t1 ) < g1 (t1 , x(t1 − τ1 (t1 ))) + g2 (t1 , x(t1 − τ2 (t1 ))) + p(t1 ), which contradicts (2.10). This contradiction implies that (2.11) holds. Case (ii). If (A1) and (A0)(2) hold, according to (2.12), we obtain 0 < g1 (t1 , x(t1 − τ1 (t1 ))) + g2 (t1 , x(t1 − τ1 (t1 ))) + p(t1 ) < g1 (t1 , x(t1 − τ1 (t1 ))) + g2 (t1 , x(t1 − τ2 (t1 ))) + p(t1 ), EJDE-2006/117 EXISTENCE AND UNIQUENESS OF PERIODIC SOLUTIONS 7 which contradicts (2.10). This contradiction implies that (2.11) holds. Case (iii). If (A2) and (A0)(1) hold, according to (2.12), we obtain g1 (t1 , x(t1 − τ1 (t1 ))) + g2 (t1 , x(t1 − τ2 (t1 ))) + p(t1 ) < g1 (t1 , x(t1 − τ1 (t1 ))) + g2 (t1 , x(t1 − τ1 (t1 ))) + p(t1 ) < 0, which contradicts (2.10). This contradiction implies that (2.11) holds. Case (iv). If (A2) and (A0)(2) hold, according to (2.12), we obtain g1 (t1 , x(t1 − τ1 (t1 ))) + g2 (t1 , x(t1 − τ2 (t1 ))) + p(t1 ) < g1 (t1 , x(t1 − τ2 (t1 ))) + g2 (t1 , x(t1 − τ2 (t1 ))) + p(t1 ) < 0, which contradicts (2.10). This contradiction implies that (2.11) holds. If (2.13) (or (2.14), or (2.15)) holds, using the methods similar to those used in Case (i) - Case (iv), we can show that (2.11) holds. This completes the proof of the Claim. Let t2 = mT + t0 , where t0 ∈ [0, T ] and m is an integer. Then, using Schwarz inequality and the inequality Z t |x(t)| = |x(t0 ) + x0 (s)ds| ≤ d + t0 Z T |x0 (s)|ds, for all t ∈ [0, T ], 0 we obtain |x|∞ = max |x(t)| ≤ d + √ t∈[0,T ] This completes the proof. T |x0 |2 . 3. Main Results Theorem 3.1. Assume that (H), (A0), (A0’) and either (A1) or (A2). Then (1.1) has a unique T -periodic solution. Proof. From Lemma 2.4, together with (H), (A0) and (A0’), it i s easy to see that (1.1) has at most one T -periodic solution. Thus, to prove Theorem 3.1, it suffices to show that (1.1) has at least one T -periodic solution. To do this, we shall apply Lemma 2.1. Firstly, we will claim that the set of all possible T -periodic solutions of (2) is bounded. Let x(t) be a T -periodic solution of equation (2). In view of (A0’)(1) and (A0’)(2), we shall consider two cases as follows. Case (i). If (A0’)(1) holds, multiplying both sides of (2) by x0 (t) and then integrating them from 0 to T , from (2.1), (2.2), (2.11), (H), (A0’)(1) and the Schwarz 8 J. XIAO,B. LIU EJDE-2006/117 inequality, we have |x0 |22 Z = T |x0 (t)|2 dt 0 T Z 0 Z 0 Bx (t − δ)x (t)dt + λ =− T g1 (t, x(t − τ1 (t)))x0 (t)dt 0 0 Z T +λ g2 (t, x(t − τ2 (t)))x0 (t)dt + λ 0 Z T p(t)x0 (t)dt 0 (g1 (t, x(t − τ1 (t))) − g1 (t, x(t)) + g1 (t, x(t)) 0 T Z 0 T Z ≤ |B||x0 |22 + |p|2 |x0 |2 + λ − g1 (t, 0))x (t)dt + λ (g2 (t, x(t − τ2 (t))) − g2 (t, x(t)) + g2 (t, x(t)) 0 T Z 0 Z 0 − g2 (t, 0))x (t)dt + λ g1 (t, 0)x (t)dt + λ 0 T g2 (t, 0)x0 (t)dt 0 T Z |B||x0 |22 + |p|2 |x0 |2 + b1 ( 1 |x(t − τ1 (t)) − x(t)|2 dt) 2 |x0 |2 + b1 |x|2 |x0 |2 0 Z + b2 ( T 1 |x(t − τ2 (t)) − x(t)|2 dt) 2 |x0 |2 + b2 |x|2 |x0 |2 0 √ + ( max |g1 (t, 0)| + max |g2 (t, 0)|) T |x0 |2 t∈[0,T ] t∈[0,T ] Z = |B||x0 |22 + |p|2 |x0 |2 + b1 ( T 1 |x(t − (τ1 (t) − K1 T )) − x(t)|2 dt) 2 |x0 |2 + b1 |x|2 |x0 |2 0 Z + b2 ( T 1 |x(t − (τ2 (t) − K2 T )) − x(t)|2 dt) 2 |x0 |2 + b2 |x|2 |x0 |2 0 √ + ( max |g1 (t, 0)| + max |g2 (t, 0)|) T |x0 |2 t∈[0,T ] t∈[0,T ] √ √ T T ≤ [|B| + b1 ( 2µ1 + ) + b2 ( 2µ2 + )]|x0 |22 + |p|2 |x0 |2 π π √ + (b1 d + b2 d + max |g1 (t, 0)| + max |g2 (t, 0)|) T |x0 |2 . t∈[0,T ] t∈[0,T ] (3.1) Now, let √ |p|2 + (b1 d + b2 d + maxt∈[0,T ] |g1 (t, 0)| + maxt∈[0,T ] |g2 (t, 0)|) T √ √ D1 = . 1 − |B| − (b1 ( 2µ1 + Tπ ) + b2 ( 2µ2 + Tπ )) In view of (2.9) and (3.1), we obtain |x0 |2 ≤ D1 , |x|∞ ≤ d + √ T D1 . (3.2) Case (ii). If (A0’)(2) holds, multiplying both sides of (2) by x0 (t − δ) and then integrating them from 0 to T , from (2.1), (2.2), (2.9), (2.11), (A0’)(2) and the EJDE-2006/117 EXISTENCE AND UNIQUENESS OF PERIODIC SOLUTIONS 9 inequality of Schwarz, we have Z T B|x0 (t − δ)|2 dt| |B||x0 |22 = | 0 T Z 0 T Z 0 g1 (t, x(t − τ1 (t)))x0 (t − δ)dt x (t − δ)x (t)dt + λ =|− 0 0 Z T +λ g2 (t, x(t − τ2 (t)))x0 (t − δ)dt + λ Z 0 T p(t)x0 (t − δ)dt| 0 ≤ |x0 |22 + |p|2 |x0 |2 + |λ T Z (g1 (t, x(t − τ1 (t))) − g1 (t, x(t)) + g1 (t, x(t)) 0 − g1 (t, 0))x0 (t − δ)dt + λ Z T (g2 (t, x(t − τ2 (t))) − g2 (t, x(t)) + g2 (t, x(t)) 0 Z 0 − g2 (t, 0))x (t − δ)dt + λ T Z 0 0 ≤ |x0 |22 Z + |p|2 |x |2 + b1 ( 0 T g1 (t, 0)x (t)dt + λ g2 (t, 0)x0 (t − δ)dt| 0 T 1 |x(t − τ1 (t)) − x(t)|2 dt) 2 |x0 |2 + b1 |x|2 |x0 |2 0 Z + b2 ( T 1 |x(t − τ2 (t)) − x(t)|2 dt) 2 |x0 |2 + b2 |x|2 |x0 |2 0 √ + ( max |g1 (t, 0)| + max |g2 (t, 0)|) T |x0 |2 t∈[0,T ] t∈[0,T ] √ √ T T ≤ [1 + b1 ( 2µ1 + ) + b2 ( 2µ2 + )]|x0 |22 + |p|2 |x0 |2 π π √ + (b1 d + b2 d + max |g1 (t, 0)| + max |g2 (t, 0)|) T |x0 |2 . t∈[0,T ] t∈[0,T ] (3.3) Now, let √ |p|2 + (b1 d + b2 d + maxt∈[0,T ] |g1 (t, 0)| + maxt∈[0,T ] |g2 (t, 0)|) T √ √ D1 = . |B| − 1 − (b1 ( 2µ1 + Tπ ) + b2 ( 2µ2 + Tπ )) In view of (2.9) and (3.3), we obtain |x0 |2 ≤ D1 , |x|∞ ≤ d + √ T D1 . (3.4) If x ∈ Ω1 = {x ∈ ker L ∩ X : N x ∈ ImL}, then there exists a constant M1 such that Z T x(t) ≡ M1 and [g1 (t, M1 ) + g2 (t, M1 ) + p(t)]dt = 0. (3.5) 0 Thus, |x(t)| ≡ |M1 | < d, √ Let M = (D1 + D1 ) T + d + 1. Set for all x(t) ∈ Ω1 . (3.6) Ω = {x|x ∈ X, |x|∞ < M }. It is easy to see from (1) and (1.3) √ that N is√L-compact on Ω. We have from (3.5), (3.6) and the fact M > max{D1 T + d, D1 T + d, d} that the conditions (1) and (2) in Lemma 2.1 hold. 10 J. XIAO,B. LIU EJDE-2006/117 Furthermore, define the continuous functions Z 1 T H1 (x, µ) = (1 − µ)x + µ · [g1 (t, x) + g2 (t, x) + p(t)]dt; µ ∈ [0 1], T 0 Z 1 T H2 (x, µ) = −(1 − µ)x + µ · [g1 (t, x) + g2 (t, x) + p(t)]dt; µ ∈ [0 1]. T 0 If (A1) holds, then xH1 (x, µ) 6= 0 for allx ∈ ∂Ω ∩ ker L. Hence, using the homotopy invariance theorem, we have Z 1 T deg{QN, Ω ∩ ker L, 0} = deg{ [g1 (t, x) + g2 (t, x) + p(t)]dt, Ω ∩ ker L, 0} T 0 = deg{x, Ω ∩ ker L, 0} = 6 0. If (A2) holds, then xH2 (x, µ) 6= 0 for all x ∈ ∂Ω∩ker L. Hence, using the homotopy invariance theorem, we obtain Z 1 T [g1 (t, x) + g2 (t, x) + p(t)]dt, Ω ∩ ker L, 0} deg{QN, Ω ∩ ker L, 0} = deg{ T 0 = deg{−x, Ω ∩ ker L, 0} = 6 0. In view of all the discussions above and Lemma 2.1, Theorem 3.1 is proved. 4. Concluding remarks Example 4.1. The first-order neutral functional differential √ √ 3 1 1 2 2 2 0 sin t) + [1 − x(t − cos2 t)] + ecos t (4.1) (x(t) + x(t − δ)) = − x(t − 8 8 64 32 64 has a unique 2π-periodic solution. 1 From (4.1), we have B = 18 , g1 (x) = − 38 x, g2 (x) = 32 [1 − x] and p(t) = ecos t . √ √ √ √ Then, µ1 = supt∈[0,T ] | 642 sin2 t| = 642 < 2π, µ2 = supt∈[0,T ] | 642 cos2 t| = 642 < 2π, 1 b1 = 38 , b2 = 32 . It is straight forward to check that all the conditions needed in Theorem 3.1 are satisfied. Therefore, (4.1) has a unique 2π-periodic solution. Remark 4.2. Equation (4.1) is a very simple version of first order NFDE. Since B 6= 0, all the results in the references and their references can not be applicable to (4.1) to obtain the existence and uniqueness of 2π-periodic solutions. This implies that the results of this paper are essentially new. Remark 4.3. By using the methods similarly to those used for (1.1), we can deal with the NFDE with multiple deviating arguments, for example (x(t) + Bx(t − δ))0 = n X gi (t, x(t − τi (t))) + p(t), (4.2) i=1 where τi (i = 1, 2, . . . , n), p : R → R and gi (i = 1, 2, . . . , n) : R × R → R are continuous functions, τi (i = 1, 2, . . . , n) and p are T -periodic, gi , i = 1, 2, . . . , n, are T -periodic in the first argument, and T > 0. One may also establish the results similarly to those in Theorem 3.1 under some minor additional assumptions on gi (t, x)(i = 1, 2, . . . , n). EJDE-2006/117 EXISTENCE AND UNIQUENESS OF PERIODIC SOLUTIONS 11 References [1] T. A. Burton, Stability and Periodic Solution of Ordinary and Functional Differential Equations, Academic Press, Orland, FL., 1985. [2] R. E. Gaines, J. Mawhin; Coincide degree and nonlinear differential equations, Lecture Notes in Math., 568, Spring-Verlag, 1977. [3] G. H. Hardy, J. E. Littlewood, G. Polya; Inequalities, London: Cambridge Univ. Press, 1964. [4] J. K. Hale, Theory of Functional Differential Equations, Spring, New York, 1977. [5] J. K. Hale, J. Mawhin; Coincide degree and periodic solutions of neutral equations, J. Differential Equations, 15(1975), 295-307. [6] V. B. Komanovskii, V. R. Nosov; Stability of Functional Differential Equations, Academic Press, London, 1986. [7] Yang Kuang, Delay Differential Equations with Applications in Population Dynamical system, Academic Press, New York, 1993. [8] Bingwen Liu and Lihong Huang, Periodic Solutions for a Class of Forced Liénard-type Equations, Acta Mathematicae Applicatae Sinica, English Series, 21(2005), 81-92. [9] Shiping Lu, Weigao Ge; On existence of periodic solutions for neutral differential equation, Nonlinear Analysis, 54(2003), 1285-1306. [10] Shiping Lu, Weigao Ge; Periodic solutions for a kind of Liéneard equations with deviating arguments, J. Math. Anal. Appl., 249(2004), 231-243. [11] Genqiang Wang, Existence of periodic solutions for second order nonlinear neutral delay equations (in Chinese), Acta Mathematica Sinica, 47 (2004), 379-384 . [12] Meirong Zhang; Periodic solutions of linear and quasilinear neutral functional differential equations, J. Math. Anal. Appl., 189(1995), 378-392. Jinsong Xiao Department of Mathematics and Computer Science, Hunan City University, Yiyang, Hunan 413000, China E-mail address: bingxiao209@yahoo.com.cn Bingwen Liu College of Mathematics and Information Science, Jiaxing University, Jiaxing, Zhejiang 314001, China E-mail address: liubw007@yahoo.com.cn